ctDNA as a Blood-based Biomarker for Immunotherapy Response

Monitoring patient response in immunotherapies is essential to optimizing treatments in the clinic. And, studies have shown that circulating tumor DNA (ctDNA) can provide an early indicator of whether or not these treatments are working. Working on this basis, researchers recently evaluated whether a blood-based biomarker detected by ddPCR could indicate how well patients with non-small cell lung cancer (NSCLC) respond to immunotherapies. Liquid biopsy returned results within two weeks instead of six weeks, when physicians routinely perform diagnostic imaging tests and evaluate disease status.

To find out more about this research and the technology that enabled it, we spoke to Jeroen Hiltermann, MD, PhD, physician and research at University Medical Center Groningen in the Netherlands. 

Could you please provide us with an overview of the work you presented at ASCO 2017?

Immunotherapy is giving patients with metastasized NSCLC a second chance at finding the right treatment for them. In fact, with new immunomodulatory drugs like the PD-1 inhibitors pembrolizumab and nivolumab, we’ve seen patients exhibit durable responses like we have never seen before. Yet, only 20% of patients respond to treatment with nivolumab, which means there is still a need to find a better predictive marker of the response to treatment.

Our current research aims to study whether ctDNA in the blood can serve as a biomarker for the response to nivolumab, which could explain why it is only effective in 20% of patients. Indeed, the preliminary data presented at ASCO which have been expanded upon since, demonstrated that ctDNA is very promising in this respect. So far, we have tested 16 patients, and close to what we expected, five have responded positively to treatment. Interestingly, all five responders exhibited a characteristic pattern in KRAS+ ctDNA levels over time: they showed an initial increase in KRAS+ ctDNA the first week after treatment commenced, followed by a steep drop by the second week, after which it became undetectable. Non-responding patients, on the other hand, showed a small decrease in KRAS+ ctDNA followed by an increase in the following weeks. This told us that KRAS+ can predict a patient’s response to nivolumab after only two weeks of treatment.

What is the value of better understanding patient response to treatment?

Biomarkers such as KRAS+ ctDNA can help determine whether a patient will respond positively to nivolumab early on in treatment. This can help an oncologist decide earlier whether monotherapy with nivolumab is sufficient to treat a patient, or whether it would be beneficial to add another therapy. In the future, as we become better at predicting patients’ responses to immunotherapy, combination therapies involving multiple drugs and modalities will become the standard treatment for metastasized NSCLC patients. 

What makes ddPCR technology well suited to this kind of project?

First, ddPCR is very sensitive. In my current study, through a liquid biopsy, ddPCR detected KRAS+ ctDNA in the base-line plasma sample of approximately 70% of patients with a KRAS+ tumor biopsy. Also, since this technology only requires a simple blood draw, liquid biopsies are easy obtainable and much more patient-friendly than rather invasive tissue biopsies. In my study, we were able to monitor ctDNA levels in our patients over multiple time points over the course of several months, which would be an impossible task if we had to rely on tissue biopsies. Patients could not undergo such invasive procedures on a repeated basis like that.

What are the next steps in developing the heightened KRAS-mutated ctDNA you detected into validated biomarker for guiding precision treatment with Nivolumab?

 We will expand our cohort by another 40 patients to demonstrate the clinical validity of this biomarker assay. However, despite the small number of samples in our pilot study, KRAS+ ctDNA appears to be a promising early predictive marker of the response to nivolumab, wherein one could potentially predict a patient’s response to treatment after only two weeks.

Jeroen Hiltermann was speaking to Jack Rudd, Senior Editor for Technology Networks.